linux-kernel - Re: [PATCH v1 04/17] media: camss: Make ISPIF subdevice optional

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <X/itZVFeM0XeV9Sx@builder.lan>
Date:   Fri, 8 Jan 2021 13:07:17 -0600
From:   Bjorn Andersson <bjorn.andersson@...aro.org>
To:     Robert Foss <robert.foss@...aro.org>
Cc:     agross@...nel.org, todor.too@...il.com, mchehab@...nel.org,
        robh+dt@...nel.org, catalin.marinas@....com, will@...nel.org,
        shawnguo@...nel.org, leoyang.li@....com, geert+renesas@...der.be,
        arnd@...db.de, Anson.Huang@....com, michael@...le.cc,
        agx@...xcpu.org, max.oss.09@...il.com,
        linux-arm-msm@...r.kernel.org, linux-media@...r.kernel.org,
        devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org,
        AngeloGioacchino Del Regno <kholk11@...il.com>,
        Andrey Konovalov <andrey.konovalov@...aro.org>,
        Tomasz Figa <tfiga@...omium.org>,
        Azam Sadiq Pasha Kapatrala Syed <akapatra@...cinc.com>,
        Sarvesh Sridutt <Sarvesh.Sridutt@...rtwirelesscompute.com>,
        Laurent Pinchart <laurent.pinchart@...asonboard.com>
Subject: Re: [PATCH v1 04/17] media: camss: Make ISPIF subdevice optional

On Fri 08 Jan 06:04 CST 2021, Robert Foss wrote:

> This driver supports multiple architecture versions of the Qualcomm ISP.
> The CAMSS architecure which this driver is name after, and with the
> introduction of this series, the Titan architecture.
> 
> The ISPIF is IP-block that is only present in the CAMSS architecture.

"is an IP-block"

> In order to support the Titan architecture, make the ISPIF an optional
> subdevice.
> 
> Signed-off-by: Robert Foss <robert.foss@...aro.org>
> ---
>  .../media/platform/qcom/camss/camss-ispif.c   | 144 ++++++++++--------
>  .../media/platform/qcom/camss/camss-ispif.h   |   3 +-
>  drivers/media/platform/qcom/camss/camss.c     | 113 +++++++++-----
>  drivers/media/platform/qcom/camss/camss.h     |   2 +-
>  4 files changed, 160 insertions(+), 102 deletions(-)
> 
> diff --git a/drivers/media/platform/qcom/camss/camss-ispif.c b/drivers/media/platform/qcom/camss/camss-ispif.c
[..]
> -int msm_ispif_subdev_init(struct ispif_device *ispif,
> +int msm_ispif_subdev_init(struct camss *camss,
>  			  const struct resources_ispif *res)
>  {
> -	struct device *dev = to_device(ispif);
> -	struct platform_device *pdev = to_platform_device(dev);
> +	struct ispif_device *ispif = camss->ispif;
> +	struct platform_device *pdev = to_platform_device(camss->dev);

It seems like several of the changes in this function is replacing
dev with camss->dev. If you retained a struct device *dev = camss->dev;
you would avoid this.

>  	struct resource *r;
>  	int i;
>  	int ret;
>  
> +	if (res == NULL && ispif == NULL)

Afaict this function is called conditional on camss->ispif != NULL, and
I don't see anything that would cause res to becomes NULL if is hasn't
been before this change.

So I think this check is unnecessary?

> +		return 0;
> +
> +	ispif->camss = camss;
> +
>  	/* Number of ISPIF lines - same as number of CSID hardware modules */
> -	if (to_camss(ispif)->version == CAMSS_8x16)
> +	if (camss->version == CAMSS_8x16)
>  		ispif->line_num = 2;
> -	else if (to_camss(ispif)->version == CAMSS_8x96 ||
> -		 to_camss(ispif)->version == CAMSS_660)
> +	else if (camss->version == CAMSS_8x96 ||
> +		 camss->version == CAMSS_660)
>  		ispif->line_num = 4;
>  	else
>  		return -EINVAL;
>  
> -	ispif->line = devm_kcalloc(dev, ispif->line_num, sizeof(*ispif->line),
> -				   GFP_KERNEL);
> +	ispif->line = devm_kcalloc(camss->dev, ispif->line_num,
> +			sizeof(*ispif->line), GFP_KERNEL);
>  	if (!ispif->line)
>  		return -ENOMEM;
>  
[..]
> @@ -1393,6 +1410,9 @@ void msm_ispif_unregister_entities(struct ispif_device *ispif)
>  {
>  	int i;
>  
> +	if (!ispif)
> +		return;

I like this, but later in the patch you make the calls to this function
conditional on ispif != NULL. You should only need one of the checks.

Regards,
Bjorn